Roles for Adenosine Ribose Hydroxyl Groups in Cyclic Adenosine 5

Aug 5, 1997 - These analogues were used to assess the effect of 2'- and 3'-hydroxyl group deletion in the adenosine ribose moiety of cADPR on the Ca2+...
23 downloads 6 Views 221KB Size
Download PDF
Biochemistry 1997, 36, 9509-9517

9509

Roles for Adenosine Ribose Hydroxyl Groups in Cyclic Adenosine 5′-Diphosphate Ribose-Mediated Ca2+ Release† Gloria A. Ashamu,‡ Jaswinder K. Sethi,§ Antony Galione,§ and Barry V. L. Potter*,‡ School of Pharmacy and Pharmacology, UniVersity of Bath, ClaVerton Down, Bath BA2 7AY, U.K., and UniVersity Department of Pharmacology, Oxford UniVersity, Mansfield Road, Oxford OX1 3QT, U.K. ReceiVed March 12, 1997; ReVised Manuscript ReceiVed May 19, 1997X

Cyclic adenosine diphosphate ribose (cADPR) is a naturally occurring and potent Ca2+mobilizing agent. Structural analogues are currently required as pharmacological tools for the investigation of this topical molecule, but modifications to date have concentrated primarily upon the purine ring. Two novel dehydroxylated analogues of cADPR have now been prepared from chemically synthesized nicotinamide adenine dinucleotide (NAD+) precursors modified in the ribose moiety linked to adenine. ADP-ribosyl cyclase of Aplysia californica catalyzed the conversion of 2′A-deoxy-NAD+ and 3′A-deoxyNAD+ into the corresponding 2′A-deoxy-cADPR and 3′A-deoxy-cADPR analogues, respectively. These analogues were used to assess the effect of 2′- and 3′-hydroxyl group deletion in the adenosine ribose moiety of cADPR on the Ca2+-releasing potential of cADPR. These compounds were found to have comparatively markedly different activities as agonists for Ca2+ mobilization in sea urchin egg homogenate. 2′A-Deoxy-cADPR is similar to cADPR, whereas 3′A-deoxy-cADPR is at least 100-fold less potent, indicating that the 3′A-hydroxyl group, but not the 2′A-hydroxyl group, is essential for calcium releasing activity. EC50 values recorded were 32 nM, 58 nM, and 5 µM for cADPR, 2′A-deoxy-cADPR, and 3′Adeoxy-cADPR, respectively. Moreover, 200 nM 2′A-deoxy-cADPR was required to desensitize the cADPRsensitive Ca2+ channel to a subsequent addition of 100 nM cADPR, but 20 µM 3′A-deoxy-cADPR was required to produce the same desensitizing effect. This is in accordance with the 100-fold lower potency exhibited by the latter analogue. To further investigate the importance of the 3′-hydroxyl group, we have also synthesized 3′A-O-methyl-cADPR, in which the 3′-hydroxyl group of adenosine has been methylated and its ability potentially to donate a hydrogen atom in a hydrogen bond has been removed. Although inactive in releasing Ca2+, 3′A-O-methyl-cADPR inhibited cADPR-induced Ca2+ release in a dose-dependent manner with an approximate IC50 value of 5 µM, whereas 3′-O-methyladenosine had no effect. This further supports the requirement of a 3′-OH group for Ca2+ releasing activity. In addition, however, it suggests that this group may not be crucial for ligand-receptor recognition. Thus, replacement of the hydrogen atom of the hydroxyl with a methyl group effects a change of activity from an agonist to an antagonist of cADPR-induced Ca2+ release. Two other analogues with modifications in the 2′ and/or 3′ positions, 3′-cADPR phosphate and 2′,3′-cyclic-cADPR phosphate, were synthesized and tested for their Ca2+-mobilizing activity in sea urchin egg homogenates. Both analogues were inactive with respect to both agonistic and antagonistic activities on the cADPR-sensitive Ca2+ release mechanism. These are the first steps toward a wider structure-activity relationship for cADPR, and this is the first study to implicate a crucial role for the adenosine ribose hydroxyl groups of cADPR in the biological activity of this cyclic nucleotide. Additionally, this is the first report of a cADPR receptor antagonist that is not modified at the 8-position of the purine ring. ABSTRACT:

Cyclic adenosine diphosphate ribose (cADPR;1 structure 1, Figure 1), a metabolite of NAD+, was first discovered during investigations of the Ins(1,4,5)P3-independent Ca2+ mobilization mechanisms in sea urchin egg preparations (Clapper et al., 1987; Dargie et al., 1989). It has been suggested to be the physiological modulator of the ryanodine receptor (Galione et al., 1993a), although there is increasing evidence that cADPR may not act directly on the ryanodine

channel but that its agonistic action is mediated by an accessory protein (Walseth et al., 1993), which has recently been suggested to be calmodulin (Lee et al., 1994a). Current evidence, therefore, is indicative of a binding site distinct from that of ryanodine. cADPR releases Ca2+ by a pathway independent of D-myo-inositol 1,4,5-trisphosphate [Ins(1,4,5)P3]. The widespread occurrence of cADPR in mammalian cells and the ubiquitous presence of the enzymes that

† We thank the Medical Research Council and the Wellcome Trust (Program Grant 045491) for financial support and the University of Bath for a Research Bursary (G.A.A.). A.G. is a Wellcome Trust Fellow and B.V.L.P. is a Lister Institute Research Professor. * To whom correspondence should be addressed: Tel 1225 826639; Fax 1225 826114; E-mail [email protected]. ‡ University of Bath. § Oxford University. X Abstract published in AdVance ACS Abstracts, July 1, 1997.

1 Abbreviations: cADPR, cyclic adenosine 5′-diphosphate ribose; NAD+, nicotinamide adenine dinucleotide; DCC, N,N-dicyclohexylcarbodiimide; DCU, dicyclohexylurea; AMP, adenosine 5′-monophosphate; KH2PO4, potassium dihydrogen phosphate; H3PO4, orthophosphoric acid; NMR, nuclear magnetic resonance; HPLC, highperformance liquid chromatography; Rt, high-performance liquid chromatography retention time; PO(OEt)3, triethyl phosphate; POCl3, phosphorus oxychloride; RFU, relative fluorescence units; TEAB, triethylammonium bicarbonate; ES, electrospray.

S0006-2960(97)00567-9 CCC: $14.00

© 1997 American Chemical Society

9510 Biochemistry, Vol. 36, No. 31, 1997

Ashamu et al.

FIGURE 1: Structures of cADPR (1), 2′A-deoxy-cADPR (9), 3′Adeoxy-cADPR (10), 3′A-OMe-cADPR (14), 3′A-P-cADPR (15), and 2′A,3′A-cyclic-P-cADPR (16).

synthesize and hydrolyze it indicate the generality of this Ca2+ release system (Rusinko et al., 1989; Galione & Sethi, 1996). A plethora of mammalian tissues have also been shown to be responsive to cADPR, e.g., cardiac myocytes (Rakovic et al., 1996), pituitary cells (Koshiyama et al., 1991), and pancreatic β-cells (Takesawa et al., 1993) among others. The presence of endogenous cADPR and its ability to mobilize Ca2+ in T-lymphocyte cell lines (Guse et al., 1995) and mouse T-lymphoma cells (Bourguignon et al., 1995) has also recently been demonstrated. Cholecystokinin octapeptide has been reported to augment the synthesis of cADPR in mammalian intestinal longitudinal smooth muscle in a concentration-dependent fashion (Kuemmerle & Makhlouf, 1995). cADPR (1) is a cyclized ADP-ribose having an N-ribosyl linkage between the anomeric carbon C1′ and the N1 of the adenine ring (Figure 1).2 The N-ribosyl linkage was originally thought to be at N6, but UV data (Kim et al., 1993), together with the crystal structure of cADPR (Lee et al., 1994b) and various NMR data (Wada et al., 1995), have confirmed that the position of cyclization is between the anomeric carbon of the terminal ribose C1′ and the N1 of adenine and that the linkage is of the β-configuration. Endogenous cADPR is synthesized from NAD+ (2) by ADP-ribosyl cyclase (Lee & Aarhus, 1991; Glick et al., 1991), with the concomitant release of nicotinamide. It is 2 Note that in the nomenclature used in this paper, H 1′, -2′, -3′, N -4′, and -5′ refer to protons on the nicotinamide ribose; HN2, -4, -5, and -6 refer to protons on the nicotinamide ring; and HA1′, -2′, -3′, -4′, and -5′ refer to protons on the adenine ribose in NAD+ and NAD+ analogues as shown in Figure 3. In cADPR and analogues, H1′, -2′, -3′, -4′, and -5′ refer to protons on the ribose that previously carried the nicotinamide moiety in the NAD+ precursor and HA1′, -2′, -3′, -4′, and -5′ refer to protons on the other ribose ring. C1′ refers to the carbon atom to which H1′ is attached. N1 and N6 refer to the nitrogen atom at positions 1 and 6 on the adenine ring, respectively, as shown in Figure 1.

FIGURE 2: Formation and metabolism of cADPR (1). cADPR is synthesized from NAD+ by ADP-ribosyl cyclase and hydrolyzed to its inactive metabolite ADP-ribose by cADPR hydrolase. CD38 is a bifunctional enzyme that catalyzes both the synthesis and degradation of cADPR.

degraded by cyclic ADP-ribose hydrolase to ADP-ribose (3). Both the substrate (NAD+) and products have no Ca2+mobilizing activity (Galione & Sethi, 1996; Lee & Aarhus, 1993). The synthesis and metabolism of cADPR is shown in Figure 2. A lymphocyte differentiating cell surface antigen CD38 (Howard et al., 1993; Summerhill et al., 1993) has also been shown to catalyze both the formation and hydrolysis of cADPR. Recent work has demonstrated that invertebrate ADP-ribosyl cyclase and CD38 may be crucial enzymes in Ca2+ signaling, as these enzymes catalyze the conversion of NAD+ and NADP+ to cADPR and NAADP+, respectively. These two different products are involved in Ca2+ mobilization from Ins(1,4,5)P3 and cADPR-insensitive Ca2+ stores in sea urchin egg homogenates (Lee & Aarhus, 1995; Aarhus et al., 1995; Genazzani & Galione, 1996). ADP-ribosyl cyclase has also been shown to mediate both nitric oxide- and cGMP-induced Ca2+ release (Galione et al., 1993b; Willmott et al., 1995; Sethi et al., 1996), thus further implicating it as a putative second messengersynthesizing enzyme involved in Ca2+ regulation. cADPR has recently been synthesized by chemical cyclization of NAD+ (Yamada et al., 1994) and also through enzymatic cyclization of N1-(5′-phosphoribosyl)ATP catalyzed by NAD+ pyrophosphorylase (Gu & Sih, 1994). However, the yields of these approaches are poor and the methods are not general. We believe that the most versatile route to chemically modified analogues as potential novel agonists, antagonists, and enzyme inhibitors is via enzymatic cyclization of NAD+ analogues catalyzed by ADP-ribosyl cyclase. We and others have demonstrated the broad

Cyclic Adenosine 5′-Diphosphate Ribose Analogues

FIGURE 3: Schematic representation of the chemoenzymatic synthesis of 2′A-deoxy-cADPR and 3′A-deoxy-cADPR. 2′A-DeoxyNAD+ (7) and 3′A-deoxy-NAD+ (8) were synthesized chemically by coupling of NMN (6) to 2′A-deoxy-AMP (4) and 3′A-deoxyAMP (5), respectively, followed by enzymatic cyclization to the corresponding cyclized structures 2′A-deoxy-cADPR (9) and 3′Adeoxy-cADPR (10), respectively.

substrate specificity for this enzyme (Ashamu et al., 1995; Walseth & Lee, 1993). Enzymatic conversion of NAD+ into cADPR was first demonstrated by using crude sea urchin egg extracts (Clapper et al., 1987). Extracts from both Lytechinus pictus and Stronglyocentrotus purpuratus were both active in producing cADPR. Routine use of Aplysia cyclase is favorable over sea urchin egg cyclase due to the higher ADP-ribosyl cyclase activity and minimal (if any) hydrolase activity and the unusual abundance of this enzyme in Aplysia californica. Our current interest focuses upon identifying the key recognition elements of cADPR that contribute to its potent Ca2+ releasing activity and also those important structural elements of NAD+ that are recognized by ADP ribosyl cyclase. Our approach in understanding the structureactivity relationship of cADPR relies primarily upon synthesis of chemically modified NAD+ analogues. We report here a chemoenzymatic synthesis of two dehydroxylated analogues of cADPR, 2′A-deoxy- (9) and 3′A-deoxy-cADPR (10), from their respective deoxy-NAD+ analogues 7 and 8 (Figures 1 and 3). A preliminary account of some of this synthetic work has already been reported (Ashamu et al., 1995). In addition, three other novel cADPR analogues with modifications in the 2′A and or 3′A positions have also been synthesized (14, 15, and 16). 3′A-O-methyl-cADPR (14) was prepared via chemical synthesis of 3′A-O-methyl-NAD+ (13), from 3′A-O-methyl-adenosine (11) (Figure 4) followed by

Biochemistry, Vol. 36, No. 31, 1997 9511

FIGURE 4: Schematic representation of the synthesis of 3′A-OMecADPR. 3′A-OMe-adenosine (11) was selectively phosphorylated to 3′A-OMe-AMP (12), followed by chemical coupling to NMN to give 3′A-OMe-NAD+ (13). Enzymatic cyclization of 3′A-OMeNAD+ yielded 3′A-OMe-cADPR (14).

enzymatic cyclization to the corresponding cyclic nucleotide. 3′A-cADPR phosphate (15) (3′A-P-cADPR) and 2′A,3′Acyclic-cADPR phosphate (16) (2′A3′A-cyclic-P-cADPR) were also synthesized. We report here an evaluation of the Ca2+mobilizing activity of these cADPR analogues and the implications for structure-activity relationships. EXPERIMENTAL PROCEDURES Materials and Methods. Nicotinamide 5′-mononucleotide (NMN), 2′A-deoxyadenosine 5′-monophosphate (2′A-deoxyAMP), 3′A-deoxyadenosine 5′-monophosphate (3′A-deoxyAMP), 3′A-O-methyladenosine (3′A-OMe-adenosine), 3′Anicotinamide adenine dinucleotide (3′A-NADP), 2′A, 3′Acyclic nicotinamide adenine dinucleotide (2′A,3′A-NADP), 2-[4-(2-hydroxyethyl)piperazin-1-yl]ethane sulfonic acid (HEPES), N,N-dicyclohexylcarbodiimide (DCC), NAD+, yeast alcohol dehydrogenase, and tris(hydroxymethyl)aminomethane (Tris) were purchased from Sigma (London). Organic solvents and other reagents were obtained from Aldrich. Pyridine was dried overnight with calcium hydride, redistilled, and stored over potassium hydroxide pellets. Aplysia ovotestis extracts containing ADP-ribosyl cyclase were prepared as previously described (Lee & Aarhus, 1991) and were used crude. The final protein concentration of the Aplysia cyclase preparation used in all cases was estimated as 10 mg/mL total protein using a Bio-Rad protein estimation kit. Fluo-3 was purchased from Molecular Probes Inc.

9512 Biochemistry, Vol. 36, No. 31, 1997 (Cambridge BioScience). All other reagents were from Sigma (London). Ion-exchange chromatography was performed on an LKB-Pharmacia medium-pressure ionexchange chromatograph using a Sepharose Q fast flow column with gradients of triethylammonium bicarbonate buffer (TEAB), pH 7.6, as eluent. TEAB (1 M) was prepared by bubbling carbon dioxide gas into 1 M triethylamine solution for ca. 6 h. HPLC was performed on a Shimadzu LC-6A chromatograph with the UV detector operating at 259 nm using a combination of a Partisil 10 µm SAX guard column (10 × 0.46 cm) and a Technicol (10 × 0.46 cm) 10 µm SAX HPLC column or using a Spherisorb 10 µm SAX (25 × 0.46 cm) column with an isocratic elution using phosphate buffer (KH2PO4), pH 3.0, at a flow rate of 1 mL/ min. 1H NMR and 31P NMR spectra were recorded on either Jeol JNM GX-270 FT NMR or Jeol EX-400 FT NMR spectrometers. Chemical shifts were measured in parts per million relative to deuteriated water (D2O) for 1H NMR and to external 85% H3PO4 for 31P NMR. For the latter, δ values are positive when downfield from this reference. J values are given in Hertz (Hz). Mass spectra were recorded at the EPSRC Mass Spectrometry Service Center at the University of Swansea and at the University of Bath. Ultraviolet (UV) absorbance was measured with a Perkin-Elmer Lambda 3 UV/vis spectrophotometer. Chemical Synthesis of NAD+ Analogues: (A) 2′A-DeoxyNAD+ (7). NMN (6) (50 mg, 149.6 µmol) and 2′A-deoxy AMP (4) (80 mg, 241.5 µmol) were dissolved in 12.5 mL of 75% (v/v) pyridine/water. N,N-Dicyclohexylcarbodiimide (DCC) (2 g, 9.7 mmol) was added to the mixture and the solution was left stirring at room temperature for 7 days (Hughes et al., 1957; Fawcett & Kaplan, 1961) (Figure 3). The reaction was quenched by pouring the mixture into 150 mL of cold water, and the suspension was left at 0 °C for 2 h to allow the DCU to precipitate. DCU was filtered off through a sintered glass funnel, and the filtrate was extracted with 3 × 50 mL aliquots of chloroform. The aqueous layer was evaporated to dryness in Vacuo, and the residue was redissolved in MilliQ water and purified on a Sepharose Q ion-exchange column, using a gradient of 0-500 mM TEAB buffer, pH 7.6. 2′A-Deoxy-NAD+ eluted as the triethylammonium (TEA) salt between 160 and 180 mM TEAB, in 19% yield, and was quantified by measuring UV absorbance at 259 nm, using an extinction coefficient of 17 800 M-1 cm-1. 1H NMR (D O, 270 MHz): δ 9.3 (H 2,s), 9.1 (H 6, d, J 2 N N ) 6.2 Hz), 8.78 (HN4, d, J ) 8 Hz), 8.3 (HA8, s), 8.1 (HN5, t, J ) 7 Hz), 8.0 (HA2, s), 6.36 (HA1′, t, J ) 7 Hz), 6.02 (HN1′, d, J ) 5.5 Hz), 4.7-4.2 (9 ribose protons, HA3′, -4′, -5′, and -5′ and HN2′, -3′, -4′ -5′, and -5′), 2.8 (HA2′, m), and 2.5 (HA2′, m). 31P NMR (D2O, 270 MHz): δ -11.08 and -10.72 (2 P, JPP ) 20 Hz). m/z (+ve ion FAB) 648 (M + H)+; m/z (-ve ion FAB) 647 (M)- and 524 [M - H - nicotinamide)]-. λmax(pH 8.3) 259 nm, 259 ) 17 800 M-1 cm-1. (B) 3′A-Deoxy-NAD+ (8). NMN (15.7 mg, 47.4 µmol) and 3′A-deoxy-AMP (5) (25 mg, 75.4 µmol) were dissolved in 6 mL of 75% aqueous pyridine. DCC (0.63 g, 3.05 mmol) was added to the solution, which was left stirring at room temperature for 7 days. The reaction was poured into 50 mL of cold distilled water and the mixture was left to stand for 2 h at 0 °C, after which the precipitate formed was filtered off. Water-insoluble impurities were extracted by shaking

Ashamu et al. the filtrate with 3 × 50 mL aliquots of chloroform. The aqueous layer was dried in Vacuo, and the residue was redissolved in MilliQ water and purified by ion-exchange chromatography, using a buffer gradient of 0-300 mM TEAB, pH 7.6. 3′A-Deoxy-NAD+ eluted between 165 and 185 mM TEAB, in 25% yield, and was quantified by measuring UV absorbance at 259 nm using an extinction coefficient of 17 800 M-1 cm-1. 1 H NMR (D2O, 270 MHz): δ 9.2 (HN2, s), 9.0 (HN6, d, J ) 6.2 Hz), 8.7 (HN4, d, J ) 7.7 Hz), 8.2 (HA8, s), 8.1 (HN5, t, J ) 7.5 Hz), 8.0 (HA2, s), 6.0 (HN1′, d, J ) 6.0 Hz), 5.9 (HA1′, d, J ) 3 Hz), 4.7-4.0 (9 H), 2.3 (HA3′, m), and 2.1 (HA3′, m). 31P NMR (D2O, 270 MHz): δ -10.26 and -10.74 (2 P, JPP ) 20 Hz). MS: m/z (+ve ion FAB) 648 (M + H)+; m/z (-ve ion FAB) 647 (M)-, 524 [M - H - nicotinamide)]-. λmax(pH 8.3) 259 nm, 259 ) 17 800 M-1 cm-1. (C) 3′A-OMe-NAD+ (13). 3′A-OMe-NAD+ was synthesized from 3′A-OMe-adenosine (11) using a two-step procedure. 3′A-OMe-Adenosine was selectively phosphorylated at the 5′ position to give 3′A-OMe-AMP (12) using a general procedure reported for nucleosides (Yoshikawa et al., 1967). 3′A-OMe-Adenosine (100 mg, 0.355 mmol) was dissolved in 2 mL of PO(OEt)3 and heated for a few seconds. The resulting solution was cooled to 0 °C and 50 µL of POCl3 was added dropwise while the temperature was maintained at 0 °C. The reaction mixture was left stirring for 2 h at room temperature under anhydrous conditions, after which the reaction was quenched by stirring in 4 mL of pyridine: water (1:3) for 30 min. The solvent was removed in Vacuo and excess PO(OEt)3 was extracted into ice cold petroleum ether 40-60 °C. The resulting monophosphate (3′A-OMeAMP) was purified by ion-exchange chromatography. 3′AOMe-AMP was obtained in 48% yield (0.17 mmol) and was then coupled to NMN to obtain 3′A-OMe-NAD+ as follows. NMN (6) (149 µmol) and 3′A-OMe-AMP (12) (170 µmol) were dissolved in 8 mL of 75% aqueous pyridine. DCC (2.0 g, 9.5 mmol) was added to the solution, which was left stirring at room temperature for 7 days. The reaction was poured into 50 mL of cold distilled water and the mixture was left to stand for 2 h at 0 °C, after which the precipitate formed was filtered off. Water-insoluble impurities were extracted by shaking the filtrate with 3 × 50 mL aliquots of chloroform. The aqueous layer was dried in Vacuo and the residue was redissolved in MilliQ water and purified by ionexchange chromatography, using a buffer gradient of 0-120 mM TEAB, pH 7.6. 3′A-OMe-NAD+ coeluted with NMN between 40 and 55 mM buffer. There was no improvement in purity by using a shallow gradient of TEAB from 0 to 70 mM. Contaminating excess NMN was removed from the desired product by treating the mixture with alkaline phosphatase for 4 h. Care was taken not to leave the solution for too long to avoid degradation of 3A′-OMe-NAD+. After 4 h most of the excess NMN had been converted to nicotinamide mononucleoside, which eluted straight off the ion-exchange column during purification of 13. The sample of 3′A-OMe-NAD+ thus obtained was 83% pure by HPLC. This material was used directly to obtain the cyclized product, as further attempts to purify it would have resulted in more material loss. 3′A-OMe-AMP: 1H NMR (D2O, 400 MHz) δ 8.3 (HA8), 7.9 (HA2, s), 5.9 (HA1′, d, J ) 5.8 Hz), 4.75 (HA2′, dd, J ) 5.5 and 5.8 Hz), 4.3 (HA3′, t), 4.05 (HA4′, m), 3.9 (HA5′,5′,

Cyclic Adenosine 5′-Diphosphate Ribose Analogues m), and 3.4 (3H, Me, s); 31P NMR (D2O, 400 MHz) δ 2.3, s, 1 P (1H-decoupled); m/z (ES) 360 [100%, (M - H)-]; λmax (pH 8.3) 259 nm 259 ) 15 300 M-1 cm-1. 3′A-OMe-NAD+: 1H NMR (D2O, 270 MHz) δ 9.4 (HN2, s), 9.2 (HN6, d, J ) 6.1 Hz), 8.7 (HN4, d, J ) 7.3 Hz), 8.5 (HA8, s), 8.35 (HA2, s), 8.2 (HN5, t), 6.15 (HN1′, d, J ) 6.5 Hz), 5.9 (HA1′, d, J ) 5.2 Hz), 4.8 (HA2′, dd, J ) 5.2 and 5.5 Hz), 4.6-4.0 (9 H), and 3.5 (3 H, s, -OMe); 31P NMR (D2O, 400 MHz) δ -11.5 and -11.8 (2 P, JPP ) 19.8 Hz); λmax(pH 8.3) 256 nm. QualitatiVe Analysis of NAD+ Analogues. The authenticity of NAD+ analogues thus synthesized was confirmed by testing for coenzyme activity (Suhaldonik et al., 1977) and the ability to form a complex with cyanide. Conversion of these NAD+ analogues into the reduced form NADH by incubation with yeast alcohol dehydrogenase in the presence of 5% v/v ethanol in 0.1 M Tris buffer, pH 9, was monitored by UV analysis. A new peak at 340 nm was recorded as a result of this conversion. Complex formation with cyanide was performed as described previously (Colowick et al., 1951). The reaction was also monitored by UV and the appearance of the new cyanide adduct was recorded at 325 nm. Enzymatic Cyclization of NAD+ and NAD+ Analogues: (A) Synthesis of cADPR (1). NAD+ (2.5 mL of 1.5 mM solution) in 25 mM HEPES, pH 6.8, was incubated with 10 µL of crude ADP-ribosyl cyclase for 10 min at room temperature. The sample was analyzed before and after enzymatic incubation by HPLC using a Partisil 10 µm SAX guard column (10 × 0.46 cm) in combination with a Technicol (10 × 0.46 cm) SAX column and isocratic elution with 0.05 M phosphate buffer, pH 3. After 10 min, the incubation mixture was diluted with 100 mL of MilliQ water and the products were purified by ion-exchange chromatography using a gradient of 0-300 mM TEAB, pH 7.6. The purity of the product obtained was verified by HPLC analysis as described above. cADPR was obtained in 60% yield and was quantified by measuring UV absorbance at 254 nm using an extinction coefficient of 14 300 M-1 cm-1 (Rusinko & Lee, 1989). Rt (incubation mixture after 10 min) 4.7 (nicotinamide), 7.8 (cADPR), and 10.3 min (NAD+). 1H NMR (D2O, 400 MHz): δ 8.9 (HA2, s), 8.2 (HA8, s), 6.0 (H1′, d, J ) 3.8 Hz), 5.85 (HA1′, d, J ) 7 Hz), 5.2 (HA2′, t, J ) 7 Hz), 4.7 (HA3′), 4.6 (H2′, H4′), 4.4 (H3′, HA5′), 4.3 (H5′), 4.2 (HA4′), 4.0 (H5′), and 3.8 (HA5′). 31P (D2O, 400 MHz): δ -11.36 and -11.84 (2 P, JPP ) 14.9 Hz). MS: m/z (+ve ion FAB) 541 (M)+, m/z (-ve ion FAB) 540 (M - H)-. (B) Synthesis of 2′A-Deoxy-cADPR (9). Enzymatic cyclization of 2′A-deoxy-NAD+ was carried out as described above for NAD+. HPLC analysis was on a Spherisorb SAX column using isocratic elution with phosphate buffer (typical chromatograms are shown in Figure 5). Ion-exchange purification of the product was carried out using a gradient of 0-300 mM TEAB, pH 7.6. Nicotinamide eluted straight off the column, followed by 2′A-deoxy-cADPR between 190 and 240 mM TEAB, in 61% yield, and finally 2′A-deoxyADPR. Pure 9 was quantified by measuring UV absorbance at 254 nm using an extinction coefficient of 14 300 M-1 cm-1. Rt (incubation mixture, Figure 4) 3.1 (nicotinamide), 4.9 (2′A-deoxy-NAD+), 7.2 (2′A-deoxy-cADPR), and 10.2 min (2′A-deoxy-ADPR). 1H NMR (D2O, 400 MHz): δ 8.9 (HA2,

Biochemistry, Vol. 36, No. 31, 1997 9513

FIGURE 5: HPLC analysis of the cyclization of 2′A-deoxy-NAD+. Conversion of 2′A-deoxy-NAD+ (7) to 2′A-deoxy-cADPR was monitored by HPLC on an anion-exchange column described in the text. (A) 2′A-deoxy-NAD+ as the starting material; (B) products obtained 10 min after incubation with crude ADP-ribosyl cyclasesnicotinamide (Nic), 2′A-deoxy-cADPR (9), and 2′A-deoxyADPR.

s), 8.2 (HA8, s), 6.4 (HA1′, t, J ) 7 Hz), 6.1 (H1′, d, J ) 5.5 Hz), 4.8 (HA3′), 4.7 (H2′, t), 4.6 (H4′), 4.4 (H3′, HA5′), 4.3 (HA5′), 4.2 (HA4′), 4.0(H5′), 3.9(HA5′), 3.2 (HA2′, m), and 2.4 (HA2′, m). 31P NMR (D2O, 400 MHz): δ -11.37 and -11.68 (2 P, JPP ) 15.8 Hz) (Figure 6). MS: m/z (+ve ion FAB) 1051 (2M + H)+ dimer and 627 (M + TEA)+; m/z (-ve ion FAB) 524 (M - H)- 1049 (2M - H)-. λmax(pH 8.3) 259 nm. (C) Synthesis of 3′A-Deoxy-cADPR (10). Enzymatic cyclization of 3′A-deoxy-NAD+, purification, and the analytical procedures used were as described for NAD+ above. 3′ADeoxy-cADPR eluted from the ion-exchange column between 165 and 185 mM TEAB in 50% yield and was quantified by measuring UV absorbance at 254 nm using an extinction coefficient of 14 300 M-1 cm-1. Rt (incubation mixture after 10 min) 4.85 (nicotinamide), 11.3 (3′A-deoxy-cADPR), and 18.3 min (3′A-deoxy-NAD+). 1H NMR (D O, 400 MHz): δ 8.7 (H 2, s), 8.4 (H 8, s), 6.0 2 A A

9514 Biochemistry, Vol. 36, No. 31, 1997

FIGURE 6: Concentration response curves for cADPR, 2′A-deoxycADPR, and 3′A- deoxy-cADPR. Assay homogenate (500 µL, 2.5%) was challenged with 5 µL of increasing concentrations of cADPR (b), 2′A-deoxy-cADPR (2), and 3′A-deoxy-cADPR (1) to give the final concentrations shown. Fluo-3 fluorescence changes were recorded and correlated to Ca2+ release following calibration of assay homogenate with exogenous applications of Ca2+. The values shown relate to means ( SD of triplicate determinations. EC50 values were obtained using curve-fitting software (KaleidaGraph) and estimated to be 32 ( 4 nM, 58 ( 5 nM, and 5 ( 1 µM for cADPR, 2′A-deoxy-cADPR, and 3′A-deoxy-cADPR, respectively.

(H1′, d, J ) 6 Hz), 5.9 (HA1′, d, J ) 3 Hz), 5.2 (HA2′, m), 4.7 (H2′), 4.5 (H4′, H5′), 4.4 (H3′), 4.2 (HA5′, HA4′), 4.0 (H5′), 3.8 (HA5′), 2.8 (HA3′), and 2.1 (HA3′). 31P NMR (D2O, 400 M Hz): δ -11.56 and -12.44 (2 P, JPP ) 15.13 Hz). MS: m/z (+ve ion FAB) 526 (M + H)+ and 627 (M + TEA)+; m/z (-ve ion FAB) 524 (M - H)- and 1049 (2M - H)-. λmax(pH 8.3) 259 nm. (D) Synthesis of 3′A-OMe-cADPR (14). 3′A-OMe-NAD+ (2 × 2.5 µmol) was dissolved in 2 mL (each) of 25 mM HEPES-NaOH buffer, pH 6.8. Aplysia cyclase (20 µL, 10 mg of protein/mL concentration) was added to each solution. The mixtures were left to incubate for 20 min, after which the solutions were combined and diluted with 100 mL of MilliQ water and the products were purified by ion-exchange chromatography using a gradient of 0-200 mM TEAB. Residual product 3′A-OMe-NAD+ eluted between 60 and 70 mM, followed by 3′-OMe-cADPR between 140 and 150 mM TEAB. Pure product was obtained in 40% (2.0 µmol) yield and was quantified by measuring UV absorbance at 254 nm using an extinction coefficient of 14 300 M-1 cm-1. HPLC retention times were measured using a 3 × 0.46 cm guard column in combination with a 10 × 0.46 cm Partisil 10 µm SAX column. Rt (incubation mixture after 20 min): 1.7 (nicotinamide), 2.5 (3′A-OMe-NAD+), and 4.4 min (3′A-OMe-cADPR). 1H NMR (D2O, 400 MHz): δ 8.9 (HA2, s), 8.3 (HA8, s), 6.0 (H1′, d, J ) 3.6 Hz), 5.9 (HA1′, d, J ) 5.5 Hz), 5.3 (HA2′, dd, J ) 5.5 and 5.2 Hz), and 4.7-3.8 (9 H, ribose protons H2′, -3′, -4′, -5′, and -5′ and HA3′, -4′, -5′, and -5′). 31P NMR (D2O, 400 MHz): δ -10.55 and -11.55 (2 P). MS: m/z (-ve ion FAB) 554 (M - H)-. λmax (pH 8.3) 256 nm. (E) Synthesis of 3′A-P-cADPR (15). 3A′- NADP (2 × 3 µmol) was dissolved in 2 mL (each) of 25 mM HEPESNaOH buffer, pH 6.8. Aplysia cyclase (20 µL, 10 mg of protein/mL concentration) was added to each solution. The mixtures were left to incubate for 20 min at room temperature

Ashamu et al. after which the solutions were combined and diluted with 100 mL of MilliQ water. The cyclic product was purified by ion-exchange chromatography using a gradient of 0-500 mM TEAB. Pure 3′A-P-cADPR eluted between 120 and 145 mM TEAB, in 54% yield (3.2 µmol), and was quantified by measuring UV absorbance at 254 nm using an extinction coefficient of 14 300 M-1 cm-1. HPLC retention times were measured using a 3 × 0.46 cm guard column in combination with a 10 × 0.46 cm Partisil 10 µm SAX column. Rt (incubation mixture after 20 min): 2.1 (nicotinamide), 15.9 (3′A-P-NAD+), and 15.9 min (3′A-P-cADPR). 1H NMR (D2O, 400 MHz): δ 8.9 (HA2, s), 8.3 (HA8, s), 6.0 (H1′, d, J ) 3.7 Hz), 5.98 (HA1′, d, J ) 7.3 Hz), 5.3 (HA2′, dd, J ) 5.0 and 7.0 Hz) and 4.6-3.9 (9 H, ribose protons H2′, -3′, -4′, -5′, and -5′ and HA3′, -4′, -5′, and -5′). 31P NMR (D2O, 400 MHz): δ +1.0 (1 P, s), -11.0 and -11.75 (2 P, JPP ) 22.6 Hz). m/z (+ve ion FAB) 723.1 (M + TEA)+ and 622 (M + H)+; m/z (-ve ion FAB) 620.1 (M - H)-. λmax (pH 8.3) 256 nm. (F) Synthesis of 2′A,3′A-Cyclic- cADPRP (16). 2′A,3′Acyclic-NADP (2 × 3 µmol) was dissolved in 2 mL (each) of 25 mM HEPES-NaOH buffer, pH 6.8. Aplysia cyclase (20 µL, 10 mg of protein/mL concentration) was added to each solution. The mixtures were left to incubate for 20 min, after which solutions were combined and diluted with 100 mL MilliQ water. The products were separated and purified by ion-exchange chromatography using a gradient of 0-400 mM TEAB. 2′A,3′A-Cyclic-NADP eluted between 180 and 190 mM, followed by 2′A,3′A-cyclic-cADPRP between 260 and 290 mM TEAB. Pure product was obtained in 45% (2.7 µmol) yield and was quantified by measuring UV absorbance at 254 nm using an extinction coefficient of 14 300 M-1 cm-1. HPLC retention times were measured using 3 × 0.46 cm guard column in combination with a 10 × 0.46 cm Partisil 10 µm SAX column. Rt (incubation mixture after 20 min): 2.1 (nicotinamide), 8.3 (2′A,3′A-cyclic-cADPRP), and 9.0 min (2′A,3′A-cyclicNADP). 1H NMR (D2O, 400 MHz): δ 8.7 (HA2, s), 8.1 (HA8, s), 6.2 (H1′, s), 5.9 (HA1′, d), 5.3 (HA2′, dd, J ) 8.0 and 6.4), 5.5 (HA3′, dd, J ) 4.3 and 5.8), and 4.6-3.9 (8 H, ribose protons H2′, -3′, -4′, -5′, and -5′, and HA4′, -5′, and -5′). 31P NMR (D2O, 400 MHz): δ +19.0, (1 P, s), -10.8, and -11.2 [2 P JPP ) 13.8 Hz (1H-decoupled)]. m/z (+ve ion FAB) 604 (M + H)+; m/z (-ve ion FAB) 602.1 (M H)-. λmax (pH 8.3) 256 nm. Assessment of Ca2+ Releasing ActiVity of Cyclic Nucleotides. Analogue-induced Ca2+ mobilization in sea urchin egg homogenates by synthetic analogues was evaluated fluorometrically. Egg homogenate (2.5%) contained the Ca2+sensitive probe Fluo-3 (3 µM). The homogenates were prepared from L. pictus eggs and diluted as previously described (Willmott et al., 1995). Km and Vmax Determination for Cyclization of NAD+ and 2′A- and 3′A-Deoxy-NAD+ by ADP-Ribosyl Cyclase. The kinetics of enzymatic conversion of the above analogues in their cyclic products were investigated using an HPLC method. The conversion was followed by measuring the area under the peak for nicotinamide, a byproduct of ADP-ribosyl cyclase activity. Final substrate concentrations between 25 and 500 µM in 25 mM HEPES buffer, pH 6.8, were used. The total volume in the reaction vessel (an eppendorf) was kept at 500 µL for each reaction. The incubation mixture contained 250 µL of substrate with 225 µL of buffer and 25

Cyclic Adenosine 5′-Diphosphate Ribose Analogues

Biochemistry, Vol. 36, No. 31, 1997 9515

µL of enzyme to make a total of 500 µL. A final protein concentration of 5 µg/mL was used in the assay for NAD+ and 2.5 µg/mL for 2A′-deoxy-NAD+ and 3′A-deoxy-NAD+. Each reaction was analyzed every minute over 6 min. The relative area for nicotinamide was measured and a graph of nicotinamide area vs time was plotted at each concentration of substrate used. The reaction was carried out in duplicate. Initial velocity was calculated from the slope of this curve and an average initial velocity obtained from the two sets of data. Km and Vmax were obtained from the Lineweaver-Burk plot by standard computer regression techniques. Km values obtained were 0.07 mM, 0.74 mM and 0.50 mM and Vmax values were 22.5, 48.0, 23.7 µmoles/mg total protein/min for NAD+, 2A′-deoxy-NAD+ and 3A′-deoxy-NAD+ respectively. RESULTS AND DISCUSSION We have shown here that a carbodiimide-based method for the synthesis of NAD+ analogues is reasonably effective. The quality of material prepared after purification can be seen from a typical HPLC trace for 7 in Figure 5A. 1HNMR spectroscopy of the cyclic products confirmed the disappearance of the nicotinamide protons after enzymatic cyclization. NMR and mass spectrometry also confirmed the integrity of the cyclized product in all cases. This indicates that analogues 7, 8, and 13 are substrates for ADPribosyl cyclase. Km values obtained for NAD+, 2′A-deoxyNAD+, and 3′A-deoxy-NAD+ (0.07, 0.74, and 0.50 mM, respectively) showed that though the enzyme appears to have a higher affinity for the natural substrate NAD+, loss of the 2A′ or 3A′ hydroxyl groups has a significant, but not dramatic, effect on the recognition and substrate activity of NAD+ by ADP-ribosyl cyclase. The 1H NMR data of cADPR agree with those previously recorded (Wada et al., 1995). Proton chemical shifts were assigned with the aid of 1H-1H COSY experiments for analogues prepared. For 9, the HA1′ signal was downfield from H1′ and appeared as a triplet - as it coupled to two HA2′ protons, contrary to the signal for 1 and 10, where H1′ is downfield from HA1′ and both appeared as singlets. Similarly, a characteristic HA2′ triplet signal at δ 5.2 for cADPR (Walseth & Lee, 1993; Wada et al., 1995) was observed. However, for 9 the two HA2′ protons resonated at a lower frequency, δ 3.2 and 2.4, as a multiplet. HA2′ in 10 appeared as a multiplet at δ 5.2. The two HA3′ protons (10) also resonated at a lower frequency - δ 2.8 and 2.1; this is consistent with the loss of the deshielding effect from the neighboring hydroxyl group. The 1H NMR chemical shifts for the protons of analogue 14 resembled those of cADPR except for the O-methyl protons that resonated at δ 3.5. 1H NMR spectra for 15 and 16 were similar to those in a preliminary report (Zhang et al., 1996). AB-type signals in the 31P NMR spectra of NAD+ and cADPR analogues support the existence of the pyrophosphate group. There is a difference of ca. 5 Hz in JAB values between NAD+ and cADPR analogues. The mass of 524, measured as the negative ion (M - H)- for 9 and 10 using FAB-MS, is consistent with deletion of an oxygen atom in both cases, compared with cADPR {540 - (M - H)-}. m/z values of 1049 (M - H)- were characteristic for both deoxy forms, showing possible formation of larger molecular weight structures. This was, interestingly, not observed in the mass spectrum of cADPR. All three compounds 1, 9, and 10

FIGURE 7: Action of analogues on cADPR-sensitive Ca2+ release. Fluorescence traces show the Ca2+ release from 500 µL of sea urchin egg homogenate (2.5%) (containing the Ca2+-sensitive fluorochrome fluo-3, 3 µM) when challenged with (A) 200 nM cADPR, (B) 200 nM 2′A-deoxy-cADPR, (C) 5 µM 3′A-deoxycADPR, and (D) 20 µM 3′A-deoxy-cADPR (final concentrations as shown). 2′A-Deoxy-cADPR behaved like authentic cADPR, as 200 nM was sufficient to desensitize to a subsequent addition of cADPR (100 nM). 3′A-Deoxy-cADPR must be present in greater amounts to produce the same desensitizing effect (C and D). The traces are from single experiments representative of three similar experiments.

showed a UV absorption maximum at 259 nm; however, there is a decrease in the UV absorption maximum to 256 nm for analogues 14, 15, and 16. An HPLC analysis of the enzymatic conversion of 2′A-deoxy-NAD+ (Figure 5B) illustrates, in addition to the desired product 9, the production of nicotinamide and 2′A-deoxy-ADPR as side products of the cyclization. 2′A-Deoxy-ADPR is produced from hydrolytic breakdown of 2′A-deoxy-cADPR and was easily removed by ion-exchange chromatography. This is in line with previous reports that suggest Aplysia cyclase does exhibit some, albeit low, cADPR hydrolase activity, although the action of other hydrolytic enzymes in the crude extracts cannot be ruled out. To investigate whether hydroxyl group deletion altered the biological activity of cADPR, we tested the Ca2+-mobilizing ability of both 2′A-deoxy-cADPR and 3′A-deoxy-cADPR. Like cADPR, the deoxy analogues were Ca2+-mobilizing agonists. However, dose-response curves differed (Figure 6). The concentrations required to produce half-maximal Ca2+ release (EC50) were estimated to be 32 ( 4 nM, 58 ( 5 nM, and 5 ( 1 µM for 1, 9, and 10, respectively. A dramatic difference is observed between the dose-response curves for 9 and 10 as shown in Figure 6. 3′A-DeoxycADPR has a higher EC50 value; hence, a higher concentration is required to release the same amount of Ca2+ as cADPR. Moreover, 200 nM 2′A-deoxy cADPR desensitized cADPR-sensitive Ca2+ release channels to subsequent addition of cADPR (100 nM), but 20 µM 3′A-deoxy-cADPR was required to produce the same effect (Figure 7). This is in agreement with the 100-fold lower potency exhibited by

9516 Biochemistry, Vol. 36, No. 31, 1997

FIGURE 8: Inhibition of cADPR-induced Ca2+ release by 3′A-OMecADPR. Experimental conditions were as described for Figure 6. Fluorescence traces show the effect of pretreating sea urchin egg homogenates (2.5%) with 3′A-OMe-cADPR on subsequent cADPRinduced Ca2+ release. While 3′A-OMe-cADPR is not itself active in releasing Ca2+, it did inhibit Ca2+ release by subsequent application of cADPR (100 nM). Concentrations of 3′A-OMecADPR used ranged from 250 nM to 10 µM as indicated in the figure. This antagonistic activity was concentration-dependent and resembled that previously reported for 8-substituted cADPR analogues, 8-NH2-cADPR and 8-Br-cADPR (Walseth & Lee, 1993). Approximately half-maximal inhibition (IC50) was produced by one treatment with 5 µM 3′-OMe-cADPR. (Note that the 0 and 500 nM traces are superimposed.)

analogue 10. Since both analogues are able to prevent cADPR-sensitive Ca2+ release, they are likely to be acting to release Ca2+ at the cADPR receptor in a similar fashion to cADPR itself. However, it is currently unclear whether this inhibitory effect is due to desensitization of cADPRsensitive receptors or to store depletion. To test whether this was due to differing rates of analogue breakdown by endogenous cADPR hydrolase, we incubated in control experiments cADPR, 2′A-deoxy-cADPR, and 3′A-deoxycADPR with sea urchin homogenates. HPLC analysis of products over 10 h showed that 3′A-deoxy-cADPR was not metabolized faster than cADPR or 2′A-cADPR. This suggested that the reduced agonistic activity of 10 is not due to decreased stability or an increased susceptibility to hydrolytic breakdown under the experimental conditions (data not shown). Removal of an oxygen atom from cADPR to give the carbocyclic cyclic aristeromycin diphosphate ribose was recently demonstrated to confer considerable metabolic resistance to this compound relative to cADPR (Bailey et al., 1996a), and deletion of the N7 purine nitrogen results in a compound with partial agonist properties (Bailey et al., 1997). Thus, the present results provide further evidence that small chemical modifications, like hydroxyl group deletion as here, can have a dramatic effect on the biological activity of a cADPR analogue. Because of the large activity difference observed between 9 and 10 we investigated further the role of the 3′A-hydroxyl group. Thus, 3′A-OMe-cADPR was synthesized. Unlike the OH moiety, the OMe group is larger and does not possess the ability to donate a proton in a potential H-bond interaction with the receptor, although it can still act as an acceptor. When tested for agonistic activity, 3′A-OMe-cADPR was not active in releasing Ca2+ in the sea urchin homogenate system (Figure 8). To investigate whether this molecule was

Ashamu et al. interacting with cADPR receptors, we tested its ability to interfere with cADPR-induced Ca2+ release. It acted, surprisingly, as an antagonist, in that cADPR-induced Ca2+ release was inhibited in a concentration-dependent manner (Figure 8). 3-O-Methyladenosine, in contrast, had no such effect up to 100 µM, nor did it act as an agonist for Ca2+ release (data not shown). It thus appears that the oxygen atom is somehow important for interaction at the receptor site, as 3′A-OMe-cADPR must interact with the receptor in order to inhibit cADPR-induced Ca2+ release in a concentration-dependent manner and yet 3′A-deoxy-cADPR is a very poor agonist. 3′A-Deoxy-cADPR can clearly neither donate nor accept a hydrogen bond at the 3′A position; hence it appears that ability of an analogue to potentially accept a hydrogen bond as in 14 may be important for receptor interaction. It is interesting that mere substitution of the hydrogen atom of the 3′A-hydroxyl group of 1 by a methyl group is sufficient to convert the ligand from an agonist to an antagonist. In contrast, analogues 15 and 16 were not active up to concentrations of 20 µM in mobilizing Ca2+ and did not inhibit cADPR-induced Ca2+ release in any way, being therefore neither agonists nor antagonists. This is consistent with the results of a previous preliminary study where 15 and 16 were tested for Ca2+ release in rat brain microsomes (Zhang et al., 1995) and found to contrast with the agonistic activity of 2′A-cADPR phosphate. While no full details of the synthesis or biology have, however, yet been reported by these authors, we provide a much fuller account in this work. 3′A-P-cADPR (15) is similar in structure to cADPR except for the presence of a phosphate group at the 3′A hydroxyl. 2′A,3′A-P-cADPR (16) also possesses a cyclic phosphate, substituting both 2′A and 3′A hydroxyls (Figure 1). Analogue 15 also has two extra negative charges and (16) has one extra negative charge at physiological pH as compared to cADPR. It is likely that a combination of these factors may result in electrostatic repulsion and steric hinderance of the analogues at the cADPR receptor site, relative to cADPR, thus preventing binding to the receptor. 2′A-P-cADPR, an analogue with a phosphate group at the 2′A position, was reported to be active at mobilizing Ca2+ from rat brain microsomes (Zhang et al., 1995; Vu et al., 1996) and permeabilized T-lymphocytes (Guse et al., 1997); however, we found it to be inactive in mobilizing Ca2+ in sea urchin egg homogenates. This is consistent with a previous report (Aarhus et al., 1995) and may reflect subtle differences in the cADPR-sensitive Ca2+ release mechanism for the ryanodine receptor of sea urchin eggs and mammalian tissue. However, we have demonstrated here that replacement of the 2′A-OH group with a hydrogen atom has little effect on the activity of cADPR, yet attachment of a phosphate group completely abolishes its activity. This inactivity of 2′A-P-cADPR in sea urchin eggs may be due to the above similar factors. It is unclear, however, at present whether the reason lies with the removal of the -OH proton by substitution or rather is the result of charge addition and/ or increase in steric volume by addition of a phosphate. In summary, our results show that 2′A-deoxy-NAD+ and 3′A-deoxy-NAD+ are substrates for ADP-ribosyl cyclase and that the 2′A and 3′A hydroxyl groups appear to play some role, but not an essential one, in the functional recognition of NAD+ by ADP-ribosyl cyclase. In cADPR, noting the results with 2′A-P-cADPR discussed above, the 2′A-OH group

Cyclic Adenosine 5′-Diphosphate Ribose Analogues appears to be unimportant, and the deoxygenated analogue exhibited very similar Ca2+ release properties to cADPR itself. In contrast, the 3′A-OH group is necessary for effective Ca2+ release activity in sea urchin egg homogenates. Loss of this latter motif resulted in a 100-fold reduction in potency. Substitution of the 3′A-hydroxyl group with a methyl group did not prevent receptor interaction but instead converted the molecule from an agonist to an antagonist. This further suggests that the 3′A-OH group is important for Ca2+ releasing activity in the sea urchin system. In addition, 3′AOMe-cADPR is the first cADPR antagonist that is not modified at the 8-position of the purine ring of cADPR. This surprising property of 3′A-OMe-cADPR therefore provides a new lead compound for the design of antagonists of cADPR-induced Ca2+ release. Furthermore, the contrasting action of 2′A-P-cADPR may offer a new group of molecules for identifying and differentiating tissue- and/or speciesspecific cADPR binding proteins. Most work to date has concentrated upon structural modification to the purine ring of cADPR (Walseth et al., 1993; Graeff et al., 1996). These results now set the stage for establishing a wider structureactivity profile for cADPR and for developing new pharmacological tools to investigate and modulate the putative signaling pathway mediated by cADPR. ACKNOWLEDGMENT We thank Dr. S. M. Fortt and Dr. M. Migaud for helpful discussions, Mr. G. Cooper and Ms. C. Armah for technical assistance. REFERENCES Aarhus, R., Graeff, R. M., Dickey, D. M., Walseth, T. F., & Lee, H. C. (1995) J. Biol. Chem. 270, 30327-30333. Ashamu, G. A., Galione, A., & Potter, B. V. L. (1995) J. Chem. Soc., Chem. Commun., 1359-1360; Corrigendum (1995) 1929. Bailey, V. C., Fortt, S. M., Summerhill, R. J., Galione, A., & Potter, B. V. L. (1996) FEBS Lett. 379, 227-230. Bailey, V. C., Sethi, J. K., Fortt, S. M., Galione, A., & Potter, B. V. L. (1997) Chem. Biol. 4, 51-61. Bourguignon, L. Y. W., Chu, A., Jin, H., & Brandt, N. R. (1995) J. Biol. Chem. 270, 17917-17922. Clapper, D. L., Walseth, T. F., Dargie, P. J., & Lee, H. C. (1987) J. Biol. Chem. 262, 9561-9568. Colowick, S. P., Kaplan, N. O., & Ciotti, M. M. (1951) J. Biol. Chem. 191, 447-459. Dargie, P. J., Agre, M. C., & Lee, H. C. (1989) Cell Regul. 1, 279-290. Fawcett, C. P., & Kaplan, N. O. (1961) J. Biol. Chem. 237, 17091715. Galione, A., McDougall, A., Busa, W. B., Willmott, N., Gillot, I., & Whitaker, M. (1993a) Science 261, 348-352. Galione, A., White, A., Willmott, N., Turner, M., Potter, B. V. L., & Watson, S. P. (1993b) Nature 365, 456-459. Galione, A., & Sethi, J. (1996) in Biochemistry of Smooth Muscle Contraction, (Barany, M., Ed.) Chapt. 23, pp 295-305, Academic Press, New York.

Biochemistry, Vol. 36, No. 31, 1997 9517 Genazzani, A., & Galione, A. (1996) Biochem. J. 315, 721-725. Glick, D. L., Hellmich, M. R., Beushausen, S., Tempst, P., Bayley, H., & Strumwasser, F. (1991) Cell Regul. 2, 211-218. Graeff, R. M., Walseth, T. F., Hill, H. K., & Lee, H. C. (1996) Biochemistry 35, 379-386. Gu, Q.-M., & Sih, C. J. (1994) J. Am. Chem. Soc. 116, 74817486. Guse, A. H., da Silva, C. P., Emmrich, F., Ashamu, G. A., Potter, B. V. L., & Mayr, G. W. (1995) J. Immunol. 155, 3353-3359. Guse, A. H., da Silva, C. P., Weber, K., Armah, C. N., Ashamu, G. A., Schulze, C., Potter, B. V. L., Mayr, G. W., & Hilz, H. (1997) Eur. J. Biochem. 245, 411-417. Howard, M., Grimaldi, J. C., Bazan, J. F., Lund, F. E., SantosArgumendo, L., Parkhouse, R. M. E., Walseth, T. F., & Lee, H. C. (1993) Science 262, 1056-1059. Hughes, N. A., Kenner, G. W., & Todd, A. R. (1957) J. Chem. Soc. 735, 3733-3738. Kim, H., Jacobson, E. L., & Jacobson, M. K. (1993) Biochem. Biophys. Res. Commun. 194, 1143-1147. Koshiyama, H., Lee, H. C., & Tashijan, A. H. (1991) J. Biol. Chem. 266, 16985-16988. Kuemmerle, J. F., & Makhlouf, G. M. (1995) J. Biol. Chem. 270, 25489-25494. Lee, H. C., & Aarhus, R. (1991) Cell Regul. 2, 203-209. Lee, H. C., & Aarhus, R. (1993) Biochim. Biophys. Acta 1164, 68-74. Lee, H. C., & Aarhus, R. (1995) J. Biol. Chem. 270, 2152-2157. Lee, H. C., Aarhus, R., Graeff, R., Gurnack, M. E., & Walseth, T. F. (1994a) Nature 370, 307-309. Lee, H. C., Aarhus, R., & Levitt, D. (1994b) Nat. Struct. Biol. 1, 143. Rakovic, S., Galione, A., Ashamu, G. A., Potter, B. V. L., & Terrar, D. A. (1996) Curr. Biol. 6, 989-996. Rusinko, N., & Lee, H. C. (1989) J. Biol. Chem. 264, 1172511731. Sethi, J. K., Empson, R. E., & Galione, A. (1996) Biochem. J. 319, 613-617. Suhadolnik, R. J., Lennon, M. B., Uematsu, T., Monahan, J. E., & Baur, R. (1977) J. Biol. Chem. 252, 4125-4133. Summerhill, R. J., Jackson, D. G., & Galione, A. (1993) FEBS Lett. 335, 281-283. Takesawa, S., Nata, K., Yonekura, H., & Okamoto, H. (1993) Science 259, 370-373. Vu, C. Q., Lu, P.-J., Chen, C. S., & Jacobson, M. K. (1996) J. Biol. Chem. 271, 4747-4754. Wada, T., Inegada, K., Arimoto, K., Tokita, K., Nishina, H., Takahashi, K., Katada, T., & Sekine, M. (1995) Nucleosides Nucleotides 14, 1301-1314. Walseth, T. F., & Lee, H. C. (1993) Biochim. Biophys. Acta 1178, 235-242. Willmott, N., Sethi, J. K., Walseth, T. F., Lee, H. C., White. A. M., & Galione, A. (1995) J. Biol. Chem. 271, 3699-3705. Yamada, S., Gu, Q.-M., & Sih, C. J. (1994) J. Am. Chem. Soc. 116, 10787-10788. Yoshikawa, M., Kato, T., & Takenishi, T. (1967) Tetrahedron Lett. 50, 5065-5068. Zhang, F.-J., Gu, Q.-M., Jing, P., & Sih, C. J. (1995) Bioorg. Med. Chem. Lett. 19, 2267-2272. BI9705671